Uncovering Bias Mechanisms in Observational Studies

Ilker Demirel, Zeshan Hussain, Piersilvio De Bartolomeis, David Sontag
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:23722-23758, 2026.

Abstract

Observational studies are a key resource for causal inference but are often affected by systematic biases. Prior work has focused mainly on detecting these biases, via sensitivity analyses and comparisons with randomized controlled trials, or mitigating them through debiasing techniques. However, there remains a lack of methodology for uncovering the underlying mechanisms driving these biases, e.g., whether due to hidden confounding or selection of participants. In this work, we show that the relationship between bias magnitude and the predictive performance of nuisance function estimators (in the observational study) can help distinguish among common sources of bias. We validate our methodology through extensive synthetic experiments and a real-world case study, demonstrating its effectiveness in revealing the mechanisms behind observed biases. Our framework offers a new lens for understanding and characterizing bias in observational studies, with practical implications for improving causal inference.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-demirel26a, title = {Uncovering Bias Mechanisms in Observational Studies}, author = {Demirel, Ilker and Hussain, Zeshan and De Bartolomeis, Piersilvio and Sontag, David}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {23722--23758}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/demirel26a/demirel26a.pdf}, url = {https://proceedings.mlr.press/v306/demirel26a.html}, abstract = {Observational studies are a key resource for causal inference but are often affected by systematic biases. Prior work has focused mainly on detecting these biases, via sensitivity analyses and comparisons with randomized controlled trials, or mitigating them through debiasing techniques. However, there remains a lack of methodology for uncovering the underlying mechanisms driving these biases, e.g., whether due to hidden confounding or selection of participants. In this work, we show that the relationship between bias magnitude and the predictive performance of nuisance function estimators (in the observational study) can help distinguish among common sources of bias. We validate our methodology through extensive synthetic experiments and a real-world case study, demonstrating its effectiveness in revealing the mechanisms behind observed biases. Our framework offers a new lens for understanding and characterizing bias in observational studies, with practical implications for improving causal inference.} }
Endnote
%0 Conference Paper %T Uncovering Bias Mechanisms in Observational Studies %A Ilker Demirel %A Zeshan Hussain %A Piersilvio De Bartolomeis %A David Sontag %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-demirel26a %I PMLR %P 23722--23758 %U https://proceedings.mlr.press/v306/demirel26a.html %V 306 %X Observational studies are a key resource for causal inference but are often affected by systematic biases. Prior work has focused mainly on detecting these biases, via sensitivity analyses and comparisons with randomized controlled trials, or mitigating them through debiasing techniques. However, there remains a lack of methodology for uncovering the underlying mechanisms driving these biases, e.g., whether due to hidden confounding or selection of participants. In this work, we show that the relationship between bias magnitude and the predictive performance of nuisance function estimators (in the observational study) can help distinguish among common sources of bias. We validate our methodology through extensive synthetic experiments and a real-world case study, demonstrating its effectiveness in revealing the mechanisms behind observed biases. Our framework offers a new lens for understanding and characterizing bias in observational studies, with practical implications for improving causal inference.
APA
Demirel, I., Hussain, Z., De Bartolomeis, P. & Sontag, D.. (2026). Uncovering Bias Mechanisms in Observational Studies. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:23722-23758 Available from https://proceedings.mlr.press/v306/demirel26a.html.

Related Material